CN113792980A - Engineering design file workload assessment method and system - Google Patents
Engineering design file workload assessment method and system Download PDFInfo
- Publication number
- CN113792980A CN113792980A CN202110947403.7A CN202110947403A CN113792980A CN 113792980 A CN113792980 A CN 113792980A CN 202110947403 A CN202110947403 A CN 202110947403A CN 113792980 A CN113792980 A CN 113792980A
- Authority
- CN
- China
- Prior art keywords
- primitive
- primitives
- workload
- engineering design
- feature vector
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/048—Activation functions
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/08—Construction
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Abstract
The invention discloses a method and a system for evaluating the workload of an engineering design file, wherein the method comprises the following steps: obtaining public graphic element information in an engineering design file with known workload; simplifying the primitive data and normalizing the primitive data into a first feature vector; taking the maximum B characteristic coefficients in each type of primitive as characteristic vectors of the type of primitive, and combining the characteristic vectors of all types of primitives into a second characteristic vector; acquiring a third feature vector according to the deviation angle of any two primitives relative to the main direction of any primitive in the two primitives; training the deep neural network model; and acquiring the workload of the engineering design file to be evaluated by adopting the trained deep neural network model. The invention can effectively represent the structural space relation and the relevant invariance characteristics of the graphic primitive graph inside the design document aiming at the characteristics of the engineering design document, and automatically complete the processing and analysis of the workload of the engineering design document in a modeling mode.
Description
Technical Field
The invention relates to the field of data processing, in particular to a method and a system for evaluating the workload of an engineering design file.
Background
In the design of an engineering design system, a designer completes related design work in a computer software drawing mode. In many cases, the performance accounting department calculates the amount of the identified component graphic information in the original drawing by means of human eye identification and manual statistics. Therefore, the efficiency of manually checking the drawings is very low, the time and the labor are consumed, and the identification accuracy is not high. Especially, under the condition that the quantity of large drawings and drawings is huge, the statistics of workload is related by means of traditional human eye identification, and the conditions that the workload statistics is inaccurate and cannot be quantized can occur.
In summary, the problems of the prior art are as follows:
(1) under the current era background of big data generation, the quantity of engineering design documents is increased sharply, the counting workload by adopting the traditional manual method is low in efficiency, and the effect of automatic and accurate evaluation is lacked.
(2) The traditional data processing method fails to mine the structural space characteristics in the design document, and the spatial transformation of the design primitives is difficult to adapt, so that the characteristic information is inaccurate. For the calculation method, the traditional method adopts simple weighting to carry out calculation, and the complex nonlinear relation between the characteristic value and the workload is difficult to describe.
Disclosure of Invention
Aiming at the defects in the prior art, the engineering design file workload assessment method and the engineering design file workload assessment system provided by the invention solve the problems of low efficiency and high difficulty in the conventional engineering design file statistics.
In order to achieve the purpose of the invention, the invention adopts the technical scheme that:
the engineering design file workload assessment method comprises the following steps:
s1, obtaining common primitive information in the engineering design file with known workload, wherein the common primitive information comprises graphic types, coordinates, layers, line types and information parameters of primitives; normalizing the known workload;
s2, in the common primitive information, only retaining data of one primitive for primitive graphs overlapped in the same type under the same layer and elements with the same coordinate and the same type in different layers, and normalizing all retained primitive data to form a first feature vector;
s3, calculating the distance between every two primitive graphs according to the coordinates in the public primitive information to obtain a distance matrix;
s4, acquiring k maximum distances related to each graphic primitive in the distance matrix, and taking the mean value of the k maximum distances as the scaling factor of the graphic primitive;
s5, classifying the scaling factors of the similar primitives, calculating the characteristic coefficients of the similar primitives according to the scaling factors of the primitives, taking the maximum B characteristic coefficients in each type of primitives as the characteristic vectors of the similar primitives, and combining the characteristic vectors of all types of primitives into a second characteristic vector;
s6, acquiring the offset angles of any two primitives relative to the main direction of any primitive in the two primitives according to the coordinates of each primitive, acquiring 720 offset angles within the range of 1-720 degrees at intervals of 1 degree, and normalizing to obtain a third feature vector;
s7, training the deep neural network model by taking the first feature vector, the second feature vector, the third feature vector and the normalized workload of the engineering design file with known workload as training sample data to obtain a trained deep neural network model;
and S8, taking the first feature vector, the second feature vector and the third feature vector of the engineering design file to be evaluated as the input of the trained deep neural network model, and taking the output of the trained deep neural network model as the workload of the engineering design file to be evaluated.
Further, the graphic types of the primitives in step S1 include lines, points, circles, arcs, swap, SOLID, polygons, equal-width lines, and text graphics; and taking the gravity center of the graphic element graph as the coordinate of the graphic element.
Further, in step S2, the primitive graphics of the same type overlapped under the same layer are specifically two primitive graphics of the same type under the same layer with the same coordinates or with coordinates within a range of 5 pixels apart.
Further, the value of k in step S4 is:
wherein m is the total number of the public graphic elements in the engineering design file; beta is a constant;indicating a rounding down.
Further, the constant β has a value of 0.01.
Further, the specific method for calculating the feature coefficients of the primitive according to the scaling factor of the primitive in step S5 includes the following sub-steps:
s5-1, for each primitive, according to the formula:
acquiring a quantity parameter a; wherein λ is a constant; dmIs the scaling factor of the primitive;represents rounding down;
s5-2, obtaining a distance values with the smallest distance between the primitive and all primitives;
s5-3, according to the formula:
obtaining the closeness degree measurement distance coefficient gamma of the graphic primitive in the whole engineering design file; wherein d isjThe j-th distance value is selected from the a distance values;
s5-4, according to the formula:
σ=2/(1+1/γ)
and obtaining the characteristic coefficient sigma of the primitive.
Further, in step S6, the specific method for obtaining the offset angle of any two primitives with respect to the principal direction of any primitive in the two primitives according to the coordinates of each primitive is as follows:
according to the formula:
obtaining a primitivei main direction θ relative to the positive x-axis directioni(ii) a Wherein (x)i,yi) Is the coordinate of the primitive i; (x)av,yav) The coordinate mean value of all the graphic primitives; pi is 180 °;
according to the formula:
acquiring the main direction theta of the primitive i and the primitive j relative to the primitive iiOffset angle psii,jFurther obtaining the offset angle of any two primitives relative to the main direction of any primitive in the two primitives; wherein (x)j,yj) Is the coordinate of primitive j.
Further, the deep neural network model in step S7 includes an input layer, a first hidden layer, a second hidden layer and an output layer; the input layers are of three types and respectively correspond to a first feature vector, a second feature vector and a third feature vector; the activation functions of the first hidden layer and the second hidden layer are both: σ (x) is 1/(1+ exp (- ∑ e)hwhxh+ b)), wherein whAn h-th weight representing the neuron at that level; x is the number ofhH input components representing the level of neurons; b is a network parameter to be trained; x represents the total input; the activation function of the output layer is a modified linear unit ReLu: phi is ai(x)=max(0,∑whxh+ b) cost function C (phi, delta) becoming (phi-delta)2And/2, wherein phi is the output of the deep neural network model, and delta is the normalized workload.
An engineering design file workload assessment system is provided, comprising:
a memory storing executable instructions; and
a processor configured to execute the executable instructions in the memory to implement the above-described method.
The invention has the beneficial effects that: the method can effectively represent the structural space relation and the relevant invariance characteristics of the graphic primitive graph inside the design document according to the characteristics of the engineering design document, automatically complete the processing and analysis of the workload of the engineering design document in a modeling mode, reduce the complicated labor cost and improve the accuracy of workload evaluation.
Drawings
FIG. 1 is a schematic flow diagram of the process;
FIG. 2 is a structural diagram of a deep neural network model front-end connection neural network in the method.
Detailed Description
The following description of the embodiments of the present invention is provided to facilitate the understanding of the present invention by those skilled in the art, but it should be understood that the present invention is not limited to the scope of the embodiments, and it will be apparent to those skilled in the art that various changes may be made without departing from the spirit and scope of the invention as defined and defined in the appended claims, and all matters produced by the invention using the inventive concept are protected.
As shown in fig. 1, the method for evaluating the workload of the engineering design file includes the following steps:
s1, obtaining common primitive information in the engineering design file with known workload, wherein the common primitive information comprises graphic types, coordinates, layers, line types and information parameters of primitives; normalizing the known workload;
s2, in the common primitive information, only retaining data of one primitive for primitive graphs overlapped in the same type under the same layer and elements with the same coordinate and the same type in different layers, and normalizing all retained primitive data to form a first feature vector;
s3, calculating the distance between every two primitive graphs according to the coordinates in the public primitive information to obtain a distance matrix;
s4, acquiring k maximum distances related to each graphic primitive in the distance matrix, and taking the mean value of the k maximum distances as the scaling factor of the graphic primitive;
s5, classifying the scaling factors of the similar primitives, calculating the characteristic coefficients of the similar primitives according to the scaling factors of the primitives, taking the maximum B characteristic coefficients in each type of primitives as the characteristic vectors of the similar primitives, and combining the characteristic vectors of all types of primitives into a second characteristic vector; the value of B is 500;
s6, acquiring the offset angles of any two primitives relative to the main direction of any primitive in the two primitives according to the coordinates of each primitive, acquiring 720 offset angles within the range of 1-720 degrees at intervals of 1 degree, and normalizing to obtain a third feature vector;
s7, training the deep neural network model by taking the first feature vector, the second feature vector, the third feature vector and the normalized workload of the engineering design file with known workload as training sample data to obtain a trained deep neural network model;
and S8, taking the first feature vector, the second feature vector and the third feature vector of the engineering design file to be evaluated as the input of the trained deep neural network model, and taking the output of the trained deep neural network model as the workload of the engineering design file to be evaluated.
The graphic types of the graphic elements in the step S1 comprise lines, points, circles, arcs, SHAPE, SOLID, polygons, isowidth lines and text graphics; and taking the gravity center of the graphic element graph as the coordinate of the graphic element.
In step S2, the primitive figures overlapped in the same type in the same layer are specifically two primitive figures with the same type and the same coordinates or the coordinates within the range of 5 pixels apart.
The value of k in step S4 is:
wherein m is the total number of the public graphic elements in the engineering design file; beta is a constant, with a value of 0.01;indicating a rounding down.
The specific method for calculating the characteristic coefficient of the primitive according to the scaling factor of the primitive in step S5 includes the following sub-steps:
s5-1, for each primitive, according to the formula:
acquiring a quantity parameter a; wherein λ is a constant, value of 0.01; dmIs the scaling factor of the primitive;represents rounding down;
s5-2, obtaining a distance values with the smallest distance between the primitive and all primitives;
s5-3, according to the formula:
obtaining the closeness degree measurement distance coefficient gamma of the graphic primitive in the whole engineering design file; wherein d isjThe j-th distance value is selected from the a distance values;
s5-4, according to the formula:
σ=2/(1+1/γ)
and obtaining the characteristic coefficient sigma of the primitive.
The specific method for acquiring the offset angle of any two primitives relative to the main direction of any primitive in the two primitives according to the coordinates of each primitive in step S6 is as follows: according to the formula:
acquiring the main direction theta of the primitive i relative to the positive direction of the x axisi(ii) a Wherein (x)i,yi) Is the coordinate of the primitive i; (x)av,yav) The coordinate mean value of all the graphic primitives; pi is 180 °;
according to the formula:
acquiring the main direction theta of the primitive i and the primitive j relative to the primitive iiOffset angle psii,jFurther obtaining the offset angle of any two primitives relative to the main direction of any primitive in the two primitives; wherein (x)j,yj) Is the coordinate of primitive j.
As shown in fig. 2, the deep neural network model in step S7 includes an input layer, a first hidden layer, a second hidden layer and an output layer; the input layers are of three types and respectively correspond to a first feature vector, a second feature vector and a third feature vector; the activation functions of the first hidden layer and the second hidden layer are both: σ (x) is 1/(1+ exp (- ∑ e)hwhxh+ b)), wherein whAn h-th weight representing the neuron at that level; x is the number ofhH input components representing the level of neurons; b is a network parameter to be trained; x represents the total input; the activation function of the output layer is a modified linear unit ReLu: phi is ai(x)=max(0,∑whxh+ b) cost function C (phi, delta) becoming (phi-delta)2And/2, wherein phi is the output of the deep neural network model, and delta is the normalized workload.
The engineering design file workload evaluation system comprises:
a memory storing executable instructions; and
a processor configured to execute the executable instructions in the memory to implement the above-described method.
In a specific implementation, although the graph is composed of primitives, the primitive graph represents the shape of the primitive since the primitive may be a "graph" having a certain shape, such as a circle, a line, an arc, and the like. The workload of the engineering design file as the training data can be confirmed by the expert of the performance assessment department and is linked with the performance amount in the design file in the wage, so the value can also be taken as the workload.
The method can iteratively update the w and b values in each layer of the deep neural network model through a back propagation algorithm. Training (train set) is carried out by dividing 40% of samples in the training samples into a training set, dividing 30% of the training samples into a verification set and dividing 30% of the training samples into a test set, and the model with better prediction capability is obtained by model training. For the engineering design file to be evaluated, the corresponding first feature vector, second feature vector and third feature vector can be obtained by adopting the same method for obtaining the first feature vector, the second feature vector and the third feature vector by the engineering design file with known workload.
In one embodiment of the present invention, the common primitive information may be counted using 9-tuple data I (c, x, y, z, l, t, p1, p2, p3), where c represents a type of a common primitive graphics class (c ═ 1, …, n), and the graphics types include lines, points, circles, arcs, SHAPE, SOLID, polygons, isolines, and text graphics. x, y, z respectively represent the coordinates of the primitive graphics, and only the x, y coordinates are considered for 2-dimensional graphics, with the z coordinate being 0. If the primitive is a simple graph, such as a circle, the center of the circle is used as the coordinate of the primitive, and if the primitive is a complex graph class, the primitive information has a plurality of coordinates, and the center of gravity of the plurality of coordinates can be calculated to be used as the coordinate of the primitive graph. l represents a layer, and t represents a line type of the graph. P1-P3 represent the parameter description of the primitive information, such as the radius of a circle, the length of a line segment, and the like, the parameter meanings of the same type of graphics are determined, the different types of parameters have different meanings, and the less than 3 information parameters are represented by 0.
In conclusion, the invention can effectively represent the structural space relation and the relevant invariance characteristics of the graphic primitive graph inside the design document aiming at the characteristics of the engineering design document, automatically complete the processing and analysis of the workload of the engineering design document in a modeling mode, reduce the complicated labor cost and improve the accuracy of workload evaluation.
Claims (9)
1. A method for evaluating the workload of an engineering design file is characterized by comprising the following steps:
s1, obtaining common primitive information in the engineering design file with known workload, wherein the common primitive information comprises graphic types, coordinates, layers, line types and information parameters of primitives; normalizing the known workload;
s2, in the common primitive information, only retaining data of one primitive for primitive graphs overlapped in the same type under the same layer and elements with the same coordinate and the same type in different layers, and normalizing all retained primitive data to form a first feature vector;
s3, calculating the distance between every two primitive graphs according to the coordinates in the public primitive information to obtain a distance matrix;
s4, acquiring k maximum distances related to each graphic primitive in the distance matrix, and taking the mean value of the k maximum distances as the scaling factor of the graphic primitive;
s5, classifying the scaling factors of the similar primitives, calculating the characteristic coefficients of the similar primitives according to the scaling factors of the primitives, taking the maximum B characteristic coefficients in each type of primitives as the characteristic vectors of the similar primitives, and combining the characteristic vectors of all types of primitives into a second characteristic vector;
s6, acquiring the offset angles of any two primitives relative to the main direction of any primitive in the two primitives according to the coordinates of each primitive, acquiring 720 offset angles within the range of 1-720 degrees at intervals of 1 degree, and normalizing to obtain a third feature vector;
s7, training the deep neural network model by taking the first feature vector, the second feature vector, the third feature vector and the normalized workload of the engineering design file with known workload as training sample data to obtain a trained deep neural network model;
and S8, taking the first feature vector, the second feature vector and the third feature vector of the engineering design file to be evaluated as the input of the trained deep neural network model, and taking the output of the trained deep neural network model as the workload of the engineering design file to be evaluated.
2. The project file workload estimation method according to claim 1, wherein the graphic types of the primitives in step S1 include line, point, circle, arc, SHAPE, SOLID, polygon, equal-width line and text graphic; and taking the gravity center of the graphic element graph as the coordinate of the graphic element.
3. The method for evaluating workload of an engineering design file according to claim 1, wherein the primitive figures overlapped in the same type in the step S2 are specifically two primitive figures with the same type and the same coordinates or the coordinates within a range of 5 pixels.
5. The project file workload estimation method according to claim 4, wherein the constant β has a value of 0.01.
6. The method for evaluating the workload of engineering design files according to claim 1, wherein the specific method for calculating the feature coefficients of the primitives according to the scaling factors thereof in step S5 comprises the following sub-steps:
s5-1, for each primitive, according to the formula:
acquiring a quantity parameter a; wherein λ is a constant; dmIs the scaling factor of the primitive;represents rounding down;
s5-2, obtaining a distance values with the smallest distance between the primitive and all primitives;
s5-3, according to the formula:
obtaining the closeness degree measurement distance coefficient gamma of the graphic primitive in the whole engineering design file; wherein d isjThe j-th distance value is selected from the a distance values;
s5-4, according to the formula:
σ=2/(1+1/γ)
and obtaining the characteristic coefficient sigma of the primitive.
7. The method for evaluating workload of engineering design files according to claim 1, wherein the specific method for obtaining the offset angle of any two primitives relative to the principal direction of any primitive in the two primitives according to the coordinates of each primitive in step S6 is as follows:
according to the formula:
acquiring the main direction theta of the primitive i relative to the positive direction of the x axisi(ii) a Wherein (x)i,yi) Is the coordinate of the primitive i; (x)av,yav) The coordinate mean value of all the graphic primitives; pi is 180 °;
according to the formula:
acquiring the main direction theta of the primitive i and the primitive j relative to the primitive iiOffset angle psii,jFurther obtaining the offset angle of any two primitives relative to the main direction of any primitive in the two primitives; wherein (x)j,yj) Is the coordinate of primitive j.
8. The project file workload assessment method according to claim 1, wherein the deep neural network model in step S7 comprises an input layer, a first hidden layer, a second hidden layer and an output layer; the input layers are of three types and respectively correspond to a first feature vector, a second feature vector and a third feature vector; the activation functions of the first hidden layer and the second hidden layer are both:wherein whAn h-th weight representing the neuron at that level; x is the number ofhH input components representing the level of neurons; b is a network parameter to be trained; x represents the total input; the activation function of the output layer is a modified linear unit ReLu: phi is ai(x)=max(0,∑whxh+ b) cost function C (phi, delta) becoming (phi-delta)2And/2, wherein phi is the output of the deep neural network model, and delta is the normalized workload.
9. An engineering design file workload assessment system, comprising:
a memory storing executable instructions; and
a processor configured to execute the executable instructions in the memory to implement the method of any one of claims 1-8.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110947403.7A CN113792980B (en) | 2021-08-18 | 2021-08-18 | Engineering design file workload assessment method and system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110947403.7A CN113792980B (en) | 2021-08-18 | 2021-08-18 | Engineering design file workload assessment method and system |
Publications (2)
Publication Number | Publication Date |
---|---|
CN113792980A true CN113792980A (en) | 2021-12-14 |
CN113792980B CN113792980B (en) | 2023-07-18 |
Family
ID=78876075
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202110947403.7A Active CN113792980B (en) | 2021-08-18 | 2021-08-18 | Engineering design file workload assessment method and system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN113792980B (en) |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5189606A (en) * | 1989-08-30 | 1993-02-23 | The United States Of America As Represented By The Secretary Of The Air Force | Totally integrated construction cost estimating, analysis, and reporting system |
US20170161606A1 (en) * | 2015-12-06 | 2017-06-08 | Beijing University Of Technology | Clustering method based on iterations of neural networks |
KR102168440B1 (en) * | 2019-12-27 | 2020-10-21 | 에스앤에스이앤지 주식회사 | A plant bidding front-end engineering and design verification method and a computer-readable recording medium recording the same |
WO2020232905A1 (en) * | 2019-05-20 | 2020-11-26 | 平安科技(深圳)有限公司 | Superobject information-based remote sensing image target extraction method, device, electronic apparatus, and medium |
US20210248812A1 (en) * | 2021-03-05 | 2021-08-12 | University Of Electronic Science And Technology Of China | Method for reconstructing a 3d object based on dynamic graph network |
-
2021
- 2021-08-18 CN CN202110947403.7A patent/CN113792980B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5189606A (en) * | 1989-08-30 | 1993-02-23 | The United States Of America As Represented By The Secretary Of The Air Force | Totally integrated construction cost estimating, analysis, and reporting system |
US20170161606A1 (en) * | 2015-12-06 | 2017-06-08 | Beijing University Of Technology | Clustering method based on iterations of neural networks |
WO2020232905A1 (en) * | 2019-05-20 | 2020-11-26 | 平安科技(深圳)有限公司 | Superobject information-based remote sensing image target extraction method, device, electronic apparatus, and medium |
KR102168440B1 (en) * | 2019-12-27 | 2020-10-21 | 에스앤에스이앤지 주식회사 | A plant bidding front-end engineering and design verification method and a computer-readable recording medium recording the same |
US20210248812A1 (en) * | 2021-03-05 | 2021-08-12 | University Of Electronic Science And Technology Of China | Method for reconstructing a 3d object based on dynamic graph network |
Also Published As
Publication number | Publication date |
---|---|
CN113792980B (en) | 2023-07-18 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Ben-Shabat et al. | DeepFit: 3D surface fitting via neural network weighted least squares | |
US11823443B2 (en) | Segmenting objects by refining shape priors | |
US8412492B2 (en) | System and method for fitting feature elements using a point-cloud of an object | |
Viswanath et al. | Dimension reduction for aerodynamic design optimization | |
Gisbrecht et al. | Linear basis-function t-SNE for fast nonlinear dimensionality reduction | |
Liu et al. | Meshing point clouds with predicted intrinsic-extrinsic ratio guidance | |
CN110674685B (en) | Human body analysis segmentation model and method based on edge information enhancement | |
Yu et al. | Data‐driven polyline simplification using a stacked autoencoder‐based deep neural network | |
US20210166131A1 (en) | Training spectral inference neural networks using bilevel optimization | |
CN115393601A (en) | Three-dimensional target detection method based on point cloud data | |
Xu et al. | Segmentation of 3D outdoor scenes using hierarchical clustering structure and perceptual grouping laws | |
WO2024078112A1 (en) | Method for intelligent recognition of ship outfitting items, and computer device | |
Arbia et al. | Contextual classification in image analysis: an assessment of accuracy of ICM | |
Yu et al. | Filling gaps of cartographic polylines by using an encoder–decoder model | |
CN113792980A (en) | Engineering design file workload assessment method and system | |
Duvall et al. | Discretization-independent surrogate modeling over complex geometries using hypernetworks and implicit representations | |
Ghiu et al. | Discrete diagonal recurrences and discrete minimal submanifolds | |
US20220156416A1 (en) | Techniques for comparing geometric styles of 3d cad objects | |
Rue et al. | Identification of partly destroyed objects using deformable templates | |
EP4002190A1 (en) | Techniques for generating subjective style comparison metrics for b-reps of 3d cad objects | |
Vetsch et al. | Neuralmeshing: Differentiable meshing of implicit neural representations | |
Choporov et al. | Optimized smoothing of discrete models of the implicitly defined geometrical objects' surfaces | |
Bründl et al. | Semantic part segmentation of spatial features via geometric deep learning for automated control cabinet assembly | |
Meßmer et al. | Robust numerical integration of embedded solids described in boundary representation | |
Crozet et al. | Fast computation of local minimal distances between CAD models for dynamics simulation |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |